¹ I hope to have some plans and photographs forthcoming for an early number of the Journal. The plan which accompanies this paper is quite rough, and, though I hope it may make it easier for the reader to follow my line of argument, it lays no claim to being final or scientific.

² Suggested in Dr. Arnold's Thucydides.

³ E.g. the southern as well as the northern entrance to the λιμήν is reduced to a manageable breadth. But we may answer that there is no διάπλους for eight or nine ships in the Sikia Channel. And it could scarcely have been four times broader than that at Boidia Koilia.

Again, by assuming, with Arnold, , that μέγίєθος (Thuc. iv. 8Google Scholar, 6) can refer, not only to length, but to circumference, we can harmonize the statement that Sphacteria was about fifteen stades in μέγєθος.

⁴ Our Pylos would be a not impossible, though inefficient, Sphacteria. Identification of the description is difficult, e.g. there is no level ground in the middle, and there are no on the west (Ib. iv. 26, 7).

⁵ Ib. iv. 5, 3, Not only could the amount of walling necessary for Hagio Nikolo by no efforts have been completed in the time, but Demosthenes could not have picked out such a position as naturally defensible. Arnold is quite misleading in the Memoir at the end of his second volume.

⁶ I went over the ground carefully, and am convinced that this is no exaggeration.

⁷ E.g. there would have been a place for the Athenians to anchor, both before the battle in the harbour (Ib. iv. 13, 3), and after the blockade had begun (ib. 26, 3).

The triremes could easily have been drawn up both on the eastern side of our Sphacteria, where the chapel now stands, and at several places on the western side. There also would have been no lack of space, and no need for taking their meals on shore by relays (ib. 26, 3, ).

Again, if the southern entrance to the was at the Sikia Channel, could it be described as (ib. 8, 6)? If our Sphacteria were Prote, the description would have been If an unnamed island, In either case the description would be unintelligible.

⁸ Prote is of course out of the question, because Thuc. iv. 26, 3 shows it was no use to the Athenians for the blockade, which it must have been if so near.

⁹ Thuc. iv. 31, 2,

¹⁰ Ib. iv. 32, 1,

¹¹ Ib. iv. 32, 2 and 3.

¹² Ib. iv. 33, 1.

¹³ Ib. iv. 32, 3 and 33, 1,

¹⁴ Ib. iv. 35, 1 and 2.

¹⁵ Ib. iv. 31, 2.

¹⁶ Adolf Holm, too, even in the English translation of his History, which has had the benefit of his corrections up to 1895, is astoundingly inadequate. He ignores not only Schliemann's researches at Pylos and Sphacteria, but even Leake's. After indulging in some a priori reflections, he sums up in these words: ‘The inference is that the details of the fighting at Pylos were, as some one has remarked, invented by Thucydides as a counterpart to Plataea, in order to show how a place ought not to be besieged’ (chap. xxiii. n. 11). As at the beginning of his note he says ‘The history of the taking of Pylos etc. has no doubt been accurately narrated by Thucydides,’ we are at a loss as to whether this piece of Müller-Strübingism is a joke or a contradiction. In any case his treatment of the facts is unscientific.

¹⁷ I hope to have photographs of all the walls which I attempt to identify published hereafter in the Journal. For their approximate position see my plan, and the key which accompanies it. As this gives full details, it will not be necessary to refer to the plan again during the course of the paper.

¹⁸ But those stationed on the north-west could certainly have seen them as they ascended the hollow, if they had turned their heads.

¹⁹ Thuc. iv. 36, 2.

²⁰ Thuc. iv. 35, 4,

²¹ Ib. iv. 36, 1.

²² Ib. iv. 35, 2, etc.

²³ Ib. iv. 36, 3.

²⁴ Ib. iv. 35, 2,

²⁵ Thuc. iv. 36, 2, etc. quoted above.

²⁶ Ib. iv. 36, 2,

²⁷ Leake, , Travels in Morea, vol. i. p. 409Google Scholar, published 1830. It must be remembered that Leake's investigations, excellent as they were, were confined to a single day, April 27, 1805, for both Pylos and Sphacteria. It is difficult to discover how far the same remark in Curtius, , Peloponnesos, vol. ii. p. 179Google Scholar, published 1852, is the result of independent work, or is merely copied from Leake. The French Expédition de Moréé, vol. i. p. 4, published 1831, notices no buildings on the north peak of Sphacteria, and, what is more important, Mr. W. G. Clark, in his not sufficiently noticed book on the Peloponnesus, published in 1858, is also silent on the point.

²⁸ I wish to lay stress on the fact that not only is Pylos out of the ordinary track of scholars in the Peloponnese, but that both in Sphacteria and Pylos itself the walls I am discussing might easily escape the notice of any one who was not making a prolonged and thorough survey of the place. Moreover, there are few spots in Greece where less systematic work has been done.

²⁹ Thuc. iv. 31, 2.

³⁰ Ib. iv. 8, 6 and 7.

³¹ Following the Admiralty Chart of 1865 as against Leake and Curtius, who say 800 feet.

³² Part of this 100 yards would only require a slight wall; the half of it immediately abutting on the Sikia Channel, a strong one.

³³ Thuc. iv. 10, 4 and 11, 1.

³⁴ Ib. iv. 11, 3.

³⁵ Ib. iv. 12, 2.

³⁶ It is important to remember the words in which Thucydides describes the guarding of the mainland side: (iv. 9, 2).

On my theory the wall marked L on the plan would of course be made higher and stronger than any of the others. And it would also be true that this was the strongest side of the place as well as the best fortified. If the wall however ran down to Boidia Koilia the land side could not possibly be called the strongest side of the place, unless the lagoon was land, and the line of eastern cliffs was also therefore

³⁷ Thuc. iv. 5, 2. They finished the work and this was possible because their aim had been to minimize trouble: (Ib. 4, 3).

A glance at Plate VIII. Fig. 2 will show the strength of my argument as to the natural frontiers of Pylos.

³⁸ I caught sight of the wall because I approached Boidia Koilia from the sea as well as from the land, sailing round the outside of Pylos. I did this because I wished to explore all the coast of Pylos. But the wall would not be seen by any one who approached by land. Most travellers land at the Sikia Channel, examine the Venetian castle, and either descend from it to Boidia Koilia to the immediate west of Nestor's cave, or, more usually, listen to the local guides when they tell them the climb is impossible, and go round to Boidia Koilia by the strip of sand which runs along the bottom of the east cliffs, by the side of the lagoon.

³⁹ This must be the wall meant by their dotted lines marked Plate VI. figures 1 and 2. In the text (vol. i. p. 4) it is called a very large ‘muraille moderne.’ But this is no argument against my theory.

Plate VIII. in the French Atlas shows that their only idea of an ancient wall, at any rate for Pylos, was a regular Hellenic wall, and the description given of this ‘muraille moderne’ is all in my favour. For it is ‘composée de quartiers de Roc, posés les uns sur les autres, comme dans les constructions Cyclopéennes; cette muraille se prolonge jusqu'à la mer, et paraît avoir été construite pour défendre le passage.’ The wall is about seven to nine feet thick. A few of the stones are Cyclopean in size. But most of them are such as two or three men could lift. And the interstices are filled up with small stones and rubble. The wall itself is perhaps seven feet high. But it takes advantage of the nature of the ground, and would present a front, at many points, of twelve or fifteen feet to an attacking force.

⁴⁰ Thuc. iv. 4, 2,

It is interesting to remark that Arnold's note on this passage, his idea of the sort of wall this description would imply, might have been written for the actual wall that now exists there. He says that the construction would resemble Cyclopean architecture, only on a smaller scale. And that the interstices of the large stones would be filled with smaller ones. This is just the sort of point where Arnold's true historical insight makes him a valuable authority to be able to quote in one's favour.

⁴¹ These remains would I think repay careful investigation. A clearing away of the sand might produce startling results.

⁴² E.g. one piece of sixteen blocks of carefully built polygonal ‘Messenian’ wall. Of the bottom row six stones can be seen, and then it loses itself in the ground. Of the second row ten perfectly preserved stones remain.

⁴³ It is what one would naturally expect that these two occupations should include the whole of the promontory. They had no need to hurry through their work of fortification. And their object was to give space for a whole city, and all the buildings that a city necessarily implied.

⁴⁴ It is in the form of an inverted pyramid.

⁴⁵ It is not that it cannot be seen from a distance, but that it looks like a part of the front wall.

⁴⁶ It is bounded on the north-west or sea side by a huge natural rock.

⁴⁷ The reservations I made above as to the certainty of my identification of the ruins on Sphacteria apply with even greater force to those on Pylos, because of the mediaeval settlement. ‘The tower that I have just discussed might wall be classed as mediaeval, if Dr. Dörpfeld's criticisms on Mr. Loring's walls at Sellasia are correct.

⁴⁸ Thuc. iv. 3, 1,

⁴⁹ Thuc. iv. 2, 2.

⁵⁰ Ib. iv. 4, 1.

⁵¹ Ib. iv. 13, 1,

⁵² On my theory of the fortifications of Pylos, this must have been the place where Demosthenes drew up his ships (Ib. iv. 9, 1). It was an excellent place for it, though of course quite unfit for the beaching of the whole Athenian fleet. The only other possible place would be on the sand-ridge by Boidia Koilia, where they could be beached, either from the sea, or, if the lagoon was open to the south, from the harbour. But this of course cannot be the place, if the north line of defence ran, as I have described it, along the line of cliff, and not near the sand-ridge and the shore of Boidia Koilia. The steep shore of the Sikia Channel, and the loose rocks towards the south-west where Brasidas failed to effect his landing, are of course in any case out of the question for the beaching of ships.

⁵³ If we accept the view that the lagoon was part of the harbour, the footpath which now runs below the east cliffs, and by which one can walk from the Sikia Channel to Boidia Koilia need not cause us any difficulty. There is no reason to suppose that it was continuous, at a time when the lagoon was part of the harbour, and its shores subject to the washing of the sea, and not yet blocked by the alluvial deposit of streams possessed of no outlet. Even if it was continuous, it would have been impossible to convey μηχαναί along it from Boidia Koilia to the extreme south-east slope. An attack there would thus still be an attack of the fleet, and not of the army, and remarks about ἀπόβασις are in point.

⁵⁴ Topographische und Hypsometrische Karte des Peloponnes, Dr.Philippson, Alfred, Berlin 1891, vol ii. p. 354Google Scholar.

⁵⁵ Thuc. iv. 11, 2.

⁵⁶ Whether this could have been done with the ships at their command placed in the ordinary, and probably, though not certainly, correct interpretation of that word, I am not concerned to argue. The breadth would be small enough to give a prima facie plausibility to a theory that a policy of blocking had been intended. The point as to the impregnable character of the blocking would be a natural afterthought of the Spartans. There is nothing in the phrase to put out of the question the translation ‘ships with their prows facing each other.’ I have however examined all the passages I can find in which the word occurs in Greek, and have to admit that usage seems to give little support to this view. For other difficult places in which the word occurs in Thuc. cf. vii. 36, 3 (bis), vii. 40, 4. His use of the word in vii. 59, 2 is in favour of the ordinary view. Herodotus, in vii. 36, when describing Xerxes' bridge over the Hellespont, is more precise in his language, and uses neither word.

⁵⁷ Arnold has some good remarks on this in the Memoir at the end of his second volume.

⁵⁸ I am doubtful to what age to assign the remains of a mole, seen by me and mentioned by most of those who have written on the subject, which run out from Pylos immediately inside the Sikia Channel in a south-east direction. At the period when it was made, the lagoon could not have been land, nor could the Sikia Channe have curved then to the north. But its existence is consistent with any theory that argues that he lagoon was part of the harbour, or any that places the southern bank of the lagoon south of the Sikia Channel.

⁵⁹ All the following measurements in fathoms and cables are taken direct from the Admiralty Chart of 1865. It may be as well to add that 10 sea cables = 1 sea mile = 6,080 feet.

⁶⁰ Thuc. iv. 31, 1,

⁶¹ Ib. iv. 31, 2,

⁶² Peloponnesos ii. p. 180.

⁶³ History of Greece, part II. chap. lii. note.

⁶⁴ I think also from some points on Sphacteria. But I unfortunately omitted to take exact notes of this point.

⁶⁵ Dr. Arnold says that this was in ‘James' Naval History,’ and apparently no one has to this day verified his reference. But not only is there no trace of such a statement in any edition of James, but there is no edition extant to which Dr. Arnold could have referred. The second edition was published in 1826, and naturally contained no account of the battle of Navarino. James died in May 1827, four months before the battle, and no demand was made for a third edition till 1837, when it was edited and continued up to date by Captain Chamier. But the second vol. of Arnold's Thucydides was published in 1832. I cannot believe that Arnold invented the statement. But where he got it from I cannot yet discover. It may have escaped me in the voluminous and unindexed pages of Marshall's Royal Naval Biography, though I have searched it under every name I could think of. Captain Burrows, R.N., Chichele Professor, Oxford, has kindly looked through some other naval authorities for me, but can find no trace of the statement. The nearest thing to it is where Sir H. Coddrington says, in a letter written from his father's ship: ‘The entrance at the south end is narrow enough to render working inconvenient for a large ship.’ See Selections from the Letters of Sir H. Coddrington, p. 18. It must be remembered that this passage does not come to much, as ‘working’ means, technically, tacking against a foul wind. This would require a good deal of space.

⁶⁶ Thuc. iv. 8, 6,

⁶⁷ Thuc. iv. 13, 2,

⁶⁸ Ib. iv. 14, 1,

⁶⁹ There was a real probability that at the last moment the Peloponnesian fleet might try to escape by one of the entrances. They had pursued this policy ever since their defeats by Phormio, e.g. Thuc. iii. 33 and iv. 8. It was the remembrance of these instances of running away, the latter of which had only occurred a few days previously, that decided the Athenian admirals to enter by both channels. Otherwise they would never have risked dividing their forces in face of superior numbers (ib. iv. 8, 2 ἑξήκοντα).

⁷⁰ Ib. iv. 13, 4,

⁷¹ Mr.Clark, W. G. (Peloponnesus, p. 221Google Scholar) has already noticed that, quite apart from the question of the southern entrance, the Sikia Channel would have been almost impossible to blockade, because of the cross fire from the Athenian Wall on Pylos.

⁷² Thuc. iv. 12, 1.

⁷³ We may well believe he would have done so if he had not fallen. It must be remembered that this happened on the first day, and probably early in that. See Thuc. iv. 12 passim and ib. 13, 1,

⁷⁴ For their difficulties see Thuc. iv. 26, 2 and 3, and ib. 27, 1.

The Spartans could perhaps have raised the blockade even after the battle in the harbour if they had not been strategically demoralized, and in an utter panic as to the safety of the garrison of Sphacteria. Their ships on the spot were still numerous (a considerable number of the sixty which were given up at the armistice, Thuc. iv. 16, 1 and 3), and quite safe when properly beached and guarded by the land forces (ib. iv. 14, 4).

⁷⁵ This view is drawn from the narrative of Thuc. iv. 13, 3 to 14, 1. Thucydides says directly that the first day that the Athenian fleet arrived the Spartans did not offer battle but remained in the harbour. That this meant beaching is a priori probable, and is directly deducible from the statement that the second day etc. For the indecision cf. etc. with This indecision can only be accounted for by the fact that there was a sudden, and only a half accepted change of policy. This policy was perhaps not one of greater daring, but of greater caution. It may have arisen from distrust of the help that could be given by the land force, and may have aimed at escape. The Athenians at any rate, as we have seen, thought it worth while to guard against this contingency.

⁷⁶ We can only congratulate ourselves that, if we accept the hypothesis of an extended north-east sand-bank across the Sikia Channel, we do not at any rate make the island much longer than it is already, and that a south-east extension opposite Neokastro does not make it longer at all.

⁷⁷ See Jowett, 's Thucydides vol. ii. notes on i. 57Google Scholar, 6, and i. 103, 1, and authorities there mentioned.